setup-cuda.sh

#!/bin/bash
# CUDA Setup script for Ursa Minor Smashed model

echo "🔥 Setting up CUDA environment for Ursa Minor Smashed model..."

# Check if NVIDIA GPU is available
if ! command -v nvidia-smi &> /dev/null; then
    echo "❌ ERROR: nvidia-smi not found. Make sure NVIDIA drivers are installed."
    exit 1
fi

echo "🔍 Checking GPU information..."
nvidia-smi

# Check CUDA version
CUDA_VERSION=$(nvidia-smi | grep "CUDA Version" | awk '{print $9}' | cut -d. -f1,2)
echo "📌 Detected CUDA Version: $CUDA_VERSION"

# Create virtual environment
echo "📦 Creating virtual environment..."
python -m venv venv-cuda

# Activate virtual environment
if [[ "$OSTYPE" == "msys" || "$OSTYPE" == "win32" ]]; then
    source venv-cuda/Scripts/activate
else
    source venv-cuda/bin/activate
fi

echo "✅ Virtual environment activated"

# Upgrade pip
pip install --upgrade pip

# Install CUDA requirements
echo "🚀 Installing CUDA requirements..."
echo "This may take a few minutes as PyTorch CUDA packages are large..."

# Detect CUDA version and install appropriate PyTorch
if [[ "$CUDA_VERSION" == "12."* ]]; then
    echo "Installing PyTorch for CUDA 12.1..."
    pip install torch torchaudio torchvision --extra-index-url https://download.pytorch.org/whl/cu121
elif [[ "$CUDA_VERSION" == "11."* ]]; then
    echo "Installing PyTorch for CUDA 11.8..."
    pip install torch torchaudio torchvision --extra-index-url https://download.pytorch.org/whl/cu118
else
    echo "⚠️  Warning: Unsupported CUDA version $CUDA_VERSION"
    echo "Installing default CUDA 11.8 PyTorch..."
    pip install torch torchaudio torchvision --extra-index-url https://download.pytorch.org/whl/cu118
fi

# Install remaining requirements
echo "📋 Installing remaining dependencies..."
pip install -r requirements-cuda.txt

# Test CUDA availability
echo "🧪 Testing CUDA setup..."
python -c "
import torch
print(f'PyTorch version: {torch.__version__}')
print(f'CUDA available: {torch.cuda.is_available()}')
if torch.cuda.is_available():
    print(f'CUDA device count: {torch.cuda.device_count()}')
    print(f'Current CUDA device: {torch.cuda.current_device()}')
    print(f'CUDA device name: {torch.cuda.get_device_name()}')
    print(f'CUDA version: {torch.version.cuda}')
else:
    print('❌ CUDA not available in PyTorch installation')
    exit(1)
"

# Verify model file exists
if [[ -f "model_optimized.pt" ]]; then
    echo "✅ Model file found: model_optimized.pt"
else
    echo "⚠️  Warning: model_optimized.pt not found in current directory"
    echo "Make sure you have the model file in the same directory as this script"
fi

echo ""
echo "🎉 CUDA setup complete!"
echo ""
echo "📖 Usage Instructions:"
echo "To activate CUDA environment:"
if [[ "$OSTYPE" == "msys" || "$OSTYPE" == "win32" ]]; then
    echo "  source venv-cuda/Scripts/activate"
else
    echo "  source venv-cuda/bin/activate"
fi
echo ""
echo "To run CUDA inference:"
echo "  python inference_cuda.py --prompt 'Your prompt here'"
echo ""
echo "To run CUDA chat:"
echo "  python chat_cuda.py"
echo ""
echo "To run CUDA benchmark:"
echo "  python benchmark_cuda.py"
echo ""
echo "📊 Test your setup:"
echo "  python -c \"import torch; print('CUDA available:', torch.cuda.is_available())\""